JP2001184242A - Cache control method and computer system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prevent any cache in a shared region to be used by another application program from being pushed out when any sharing is not generated. SOLUTION: An exclusive cache region is ensured in each user space, and a shared cache region is ensured in a system space, and data read when the program of the user space (for example, B) performs access to an auxiliary memory unit are stored in the exclusive cache region of a user space B when the data are not shared, and when the program of the other user space (for example, A) performs access to the auxiliary memory unit, a cache block management table 407 is retrieved by a retrieving means 409, and when the accessed data are stored in the exclusive cache region of the user space B, the data stored in the exclusive cache region of the user space B are copied to the shared cache region by a copy means 410, and the input and output of the data for the shared cache region is operated.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a cache control method and a computer system for improving the apparent access speed by reducing the number of accesses to a storage device having a lower access speed than a main storage device such as a hard disk device. .

[0002]

2. Description of the Related Art FIG. 18 shows an example of a computer system including a cache control unit realized based on a conventional cache control method. The computer system shown in FIG. 18 has a main storage device 1801 and an auxiliary storage device 1802 for storing data to be handled. The main storage device 1801 is a system space 180 for operating an operating system.
3 and a user space 1804 for operating a user application. When the application program 1805 makes an access request to a data block in the auxiliary storage device 1802 using a system call 1806, the cache control unit 1807 uses the cache block management table 1808 and its search means 1809 to
Check whether the requested data block exists in main storage. If the data block exists, the data block in the main memory called the cache block 1810 is accessed using the input / output unit 1811, and the process ends. If the data block does not exist in the main storage, the replacement unit 18 using the LRU queue 1812
13 reads the data block into the cache block 1810 to be replaced, updates the cache block management table 1808, and then accesses the cache block.

In general, an LRU (Least) is used as an algorithm for efficiently managing a limited cache area.
(Recently Used).
This algorithm uses the property that data that has been accessed recently is more likely to be accessed again, and replaces the oldest data with the oldest access time when storing new data. At this time, an LRU queue is used as a data structure indicating the new access time.
The LRU queue used for cache control uses the address information of the cache block as its element, and the closer to the head, the more recent the access time to the data block is. If there is data access, the element corresponding to the data is Moved to the head of the queue.

As literature relating to the prior art of the cache control method, for example, published by Kyoritsu Shuppan Co., Ltd. in June 1991, written by Maurice J. Bach, translated by Sakamoto Bun, Tada Yoshikatsu and Murai Jun, "Design of UNIX Kernel" There is.

[0005]

In the above-described cache control unit, a specific application program may occupy a cache area by continuously inputting / outputting data of a large size. When the cache area is occupied, other application programs lose data from the cache, and the performance of accessing the disk device is reduced. To prevent the above-mentioned occupation of the cache area, see Japanese Patent Application Laid-Open No. 8-147218.
The "cache control device" proposes a method in which a dedicated area is provided for each processing means and used together with a shared area.

However, the above invention does not consider a sharing method when a plurality of processing units access the same data block. According to the method proposed by the above invention, when a certain data block is cached in the dedicated area A by the processing means A and another processing means B requests the same data block, the search means will only provide the dedicated area B and the shared area. Do not search. Therefore, even if the data block in the dedicated area A has been updated by the processing unit A, the processing unit B reads the data from the disk device, and if the update result is not reflected on the disk device, the processing unit B I can't know that it has been updated. If the processing means B also updates the cache block in the dedicated area B in this state, the updated data that has been reflected earlier in the disk device is overwritten by the data that has been reflected later, and becomes invalid. Therefore, this method has a problem that data consistency cannot be guaranteed when sharing occurs. An object of the present invention is to prevent flushing of a shared area cache used by another application program when sharing does not occur. Another object of the present invention is to enable the same data block to be shared by a plurality of processing means.

[0007]

In order to solve the above-mentioned problems, the present invention comprises an auxiliary storage device, and uses a main storage device in a system space where an operating system operates and a user space where an application program runs. A cache control method in a computer system, wherein a dedicated cache area for executing an application program is secured in the user space, a shared cache area is secured in the system space, and when an application program accesses the auxiliary storage device, The data in the auxiliary storage device is temporarily stored in the dedicated cache area, and when an access request is issued from another application program to the same data as the temporarily stored data in the dedicated cache area, the temporary Dedicated card stored in And so as to copy the data Gerhard region in the shared cache area.

In the case where an access request is issued from a plurality of application programs to the same data as the temporarily stored data in the dedicated cache area, and all of the access requests are read requests, Is copied to each of the dedicated cache areas for executing the other application programs.

Also, in a computer system having an auxiliary storage device and using the main storage device separately in a system space in which an operating system operates and a user space in which an application program operates, the computer system is for executing an application program in the user space. Means for reserving a dedicated cache area, means for reserving a shared cache area in the system space, and temporarily storing data read from the auxiliary storage device in the dedicated cache area when an application program accesses the auxiliary storage device. Means for storing, in a case where an access request is issued from another application program to the same data as the temporarily stored data of the dedicated cache area, the temporarily stored data of the dedicated cache area is stored in the shared cache. And to have a means for copying the shoe region.

[0010] Further, when an access request is issued from a plurality of application programs to the same data as the temporarily stored data in the dedicated cache area, and all the access requests are read requests, Means for copying the data of the dedicated cache area stored in the dedicated cache area to each of the dedicated cache areas for executing the other application programs.

[0011]

DESCRIPTION OF THE PREFERRED EMBODIMENTS A first embodiment of the present invention will be described. FIG. 1 shows a configuration related to cache control in a computer system, and includes a main storage device 101 and an auxiliary storage device 102 for storing data. Main storage device 1
Reference numeral 01 denotes a system space 103 in which an operating system operates and a plurality of user spaces 105 in which application programs 104 operate. An interface called a system call 106 is installed in the system space 103, and the application program 10
4 can use only the functions provided by the system call 106 among the functions realized in the system space 103. In addition to the cache control unit 1 in the system space 103,
07, the shared cache area 108, the space control unit 109,
A task control unit 110 and the like are provided. The cache control unit 107 according to the present invention includes a cache area management table 111, a cache block management table 112, a search unit 113, a copy unit 114, and the like.
The individual components will be described later in detail. On the other hand, a dedicated cache area 115 is secured in the user space 105, and the application program 104 in the space is used for storing a data cache.

FIG. 2 is a diagram for explaining a process in which a program in the user space A 201 generates a user space B 203 using the system call 202. User space A
When the program 204 of 201 designates the user space B and requests space activation, the system call 202 generates and activates the user space B 203 by the space control unit 205. At the same time, using the cache control unit 206,
A dedicated cache area 207 of the user space B203 is secured at a specific size. As a method of determining the size to be secured in the above processing, there are a method specified by a user when using a system call for activating a space and a method in which a system standard value is determined. A cache block area 208 is reserved in the dedicated cache area 207 of each user space for storing data blocks. An LRU queue 209 is provided in the dedicated cache area 207 to determine the priority of replacing the cache block.
In the present embodiment, the method using the LRU queue is described. However, any other management method that efficiently uses a limited cache area can be realized. The cache control unit 206 registers the area information in the cache area management table 211 using the area registration unit 210 when securing the dedicated cache area 207. Further, in the space end system call, the cache control unit 206 uses the flushing unit 212 to reflect in the auxiliary storage device 213 everything updated in the cache in the dedicated cache area of the space. Then, the cache control unit 20
6 deletes the area information from the cache area management table 211 using the registration deletion unit 214. Finally, the space is ended by the space control unit 205.

FIG. 3 shows the structure of the cache area management table 211. The cache area management table includes an address space identifier 301, a start address 302 of the dedicated cache area, an area size 303 of the dedicated cache area, and a pointer 304 indicating the LRU queue of the dedicated cache area.
And stores information on a cache area secured in the user space indicated by the space identifier 301. One of the entries relates to a shared cache area in the system space, and the other entry relates to a dedicated cache area in the user space. In this case, in the entry relating to the shared cache area, the start address of the dedicated cache area in FIG. 3 is the start address of the shared cache area, and the size of the dedicated cache area is the size of the shared cache area.

FIG. 4 is a diagram for explaining functions related to input / output processing. The application program 402 in the user space B401 calls an input / output system call 404 to perform input / output with respect to the auxiliary storage device 403. The input / output system call 404 is the cache control unit 4
The search unit 406 of step 05 is used to check where the requested block is cached. Here, the search unit 406 searches the cache block management table 407 of the cache control unit 405 at high speed.

As shown in FIG. 5, the entries of the cache block management table 407 are composed of a device number 501, a physical block address 502, an address space identifier 503, and a cache block address 504. Among them, the device number 501 and the physical block address 502 are information serving as search keys for the table, and the search means 406 specifies these two pieces of information to specify an entry.

There is a relatively high-speed implementation method of the search means 406 using a hash. In the case of this embodiment, the above 2
A hash function having two pieces of information as input is prepared, and an address where an entry of the cache block management table is stored is obtained based on the value. If it is determined in the above search process that the data is not cached, the cache control unit 405
First selects either the dedicated cache area 411 or the shared cache area 412 as an area to be cached using the area selection means 408.

The data block is cached using the replacement means 409 for the cache block determined based on the LRU queue 413 in the selected area. In this embodiment, since the LRU queue is owned for each space, LR
The entry of the U queue has a structure as shown in FIG. In addition to pointers (601 and 602) indicating previous and next entries and address information 603 of the cache block, the pointer has a pointer 604 indicating an entry of the cache block management table.

The replacement means reflects the replacement of the old cache block in the cache block management table by using the pointer. When the data block is cached, the cache control unit performs input / output processing using the input / output unit 414.

If it is found by the search process that the cache is already cached in the dedicated cache area 415 of the user space A, the cache control unit copies the cache block to the shared cache area 412 by using the copy means 410. . In this embodiment, the cache control unit uses the function of the privileged task control unit 416 to activate a processing execution unit (hereinafter, referred to as a privileged task) that can use a privileged instruction for accessing an address in a different space. Thus, the copy unit 410 is realized.

FIG. 7 shows the flow of processing at the time of input / output of the cache control unit 405. In step 701, a search process using a search unit is performed. Here, the entry of the cache block management table corresponding to the requested data block is searched. The determination unit 702 determines from the result of step 701 whether a data block exists in main storage. If it is not cached, it is necessary to cache the disk block, but if it is known that the disk block is shared, it should be cached in the shared cache area in advance. The determination unit 703 determines whether the data block should be shared using the area selection unit.

Whether or not a certain data block is shared often depends on the file. Therefore, the cache area selecting means 408 in this embodiment is shared by a flag which is set in advance for the file to which the data block belongs. Determine if it should. There are two ways to set a flag in a file: one is to explicitly specify the flag by the user, and the other is not. The method specified by the user requires an extension of the system call. As a method of realizing the area selecting means that does not extend the system call, there is a method of determining whether or not the file should be shared depending on the type of the file. For example, consider a system in which directory information used for searching for a file is handled in the same way as a file, and the directory has a hierarchical structure. When the system is applied to such a system, the directory located at a higher level is likely to share access, so that it should be cached in a shared area in advance. Further, when a file is shared, the fact that the file has been shared is stored as history information, and it can be realized by determining whether to share the file based on the information.

In accordance with the judgment of the judging unit 703, cache is performed in the area selected by the area selecting means by using the replacing means in 704 and 705. In step 706, an entry in the cache block management table is added, and the entry is exclusively locked. Exclusive lock here means that while you own the lock, other processes that try to acquire the lock wait, and if the lock fails, the process that holds the lock releases the lock. The process is in a waiting state until it is released. When the lock is secured, step 7
At 07, input / output processing is performed on the cache block using the input / output means. Finally, in step 708, the entry is unlocked, and the process ends.

FIG. 8 shows the processing performed by the cache control unit when the requested data block has already been cached by the determination unit 702. First, in step 801,
At step 701, the entry of the cache block management table obtained by using the search means is exclusively locked. The determination unit 802 determines whether the cached space is a space that can be accessed by a task that executes the request source application. Here, the accessible space is
That is, it is either a dedicated cache area in the space of the request source itself or a shared cache area in the system space, and the space other than these spaces is a space that cannot be accessed. If it is an accessible area, input / output processing is directly performed on the cache block in step 707, and
In step 708, the entry is unlocked. If the area is not an accessible area, copying is performed to the shared cache area using the copying means in step 803, input / output processing is performed on the shared cache block in step 707, and the entry lock is released in step 708. . In the above copy processing, since the application itself requesting the copy of the cache block cannot perform the copy processing directly because the address space is different,
Start the privileged task and request processing.

FIG. 9 shows the flow of processing of the part where the cache control unit starts the privileged task. Step 90
After activating the privileged task in step 1, the timer is activated in step 902 and then waiting for a response from the privileged task 90.
3. The process of the cache control unit resumes the process by releasing the wait state by the response from the privileged task or by the interruption 904 from the timer. In the determination unit 905,
If it is confirmed that the data has been correctly copied to the shared cache area, input / output processing is performed on the cache block in step 906. When the timer interrupt 904 occurs, the determination unit 905 may determine that the copy is not correct. In this case, since the privileged task may have terminated abnormally, the disk block is read into the shared cache area by itself using the replacing means in step 907, cached, and input / output processing is performed in step 906.

FIG. 10 shows a flow of processing performed by a privileged task as an embodiment of the copying means. In step 1001, copying is performed to the shared cache area, and in step 1002, the space identifier and the cache address in the cache block management table are changed to those of the shared cache area.
In step 1003, the LR of the copy source cache block
The U queue entry is moved to the end of the LRU queue and is set as a replacement candidate. Finally, in step 1004, the requesting task is notified that copying has been completed normally.

FIG. 11 illustrates the flow of processing performed by the replacing means. In step 1101, the last entry of the LRU queue is selected as a replacement target.
In step 2, the entry in the cache block management table corresponding to the entry to be replaced is deleted. When the replacement target release processing ends, the data block is read into the cache block address in step 1103, and an entry in the cache block management table corresponding to the read data block is added in step 1104.
In step 1105, the LRU entry is updated and moved to the head of the LRU queue.

By realizing the above-described table structure and processing means, caching can be performed in the user space, and the influence on the shared cache area is reduced. This also allows the user to create an application program that does not affect the disk access performance of other application programs. Further, when the same data block is accessed, data can be shared from a plurality of applications, and the result of accessing the cache block can be correctly reflected on the auxiliary storage device.

As an embodiment of the copying means 410 in FIG. 4, in a system having an inter-space access function,
The function can be realized by using the function in place of the privileged task control unit 416. In this case, since the processing of FIGS. 9 and 10 can be simplified as one processing,
0 can be improved. Second embodiment of the present invention
An embodiment of the cache control method according to the embodiment will be described below. The entry of the cache block management table in the present embodiment is configured with information as shown in FIG. The device number 1201 and the physical block address 1202 are the same as those in the embodiment of the first aspect, and are information serving as search keys. Shared cache block address 120
No. 3 stores the address of the data block indicated by the entry when the data block is cached in the shared cache area. The value is 0 when not cached in the shared cache area. In the second embodiment, since one cache block can be cached in a dedicated cache area held by a plurality of user spaces, a plurality of cache locations are associated with entries in the cache block management table. For this purpose, the entry has a pointer 1204 to the address list of the dedicated cache area. As shown in FIG. 13, the element structure of the address list has information of an address space identifier 1301 and a cache block address 1302, and further has pointers (1303, 1304) for forming a bidirectional list.

The input / output processing in this embodiment is the same as that in the first embodiment.
It is necessary to change the processing when the cache already exists in the input / output processing in. Although the above processing is shown in FIG. 8, it can be realized by replacing that part with the processing shown in FIG. The determination unit 1401 determines whether the request for the process being executed is a write request. In the case of a write request, the entry in the cache block management table is exclusively locked in step 1402, and in the case of a read request, the entry is shared locked in step 1403. Here, the process for performing the shared lock is in a wait state when the exclusive lock is secured by another process, but the shared lock can be secured for the shared lock by the separate process at the same time. In this embodiment, a shared lock is used to indicate that a plurality of input processes can be performed simultaneously. However, in a system that does not provide a shared lock, it is possible to realize only an exclusive lock.

The determination unit 1404 determines whether a cache exists in the shared cache area, that is, whether the shared cache block address of the entry is not 0. If it exists, the input / output processing is performed on the cache block in the shared cache area to release the lock, and the processing ends.

If they are not shared, the judging unit 1405 judges whether they are cached in their own space. If cached in its own space, the determination unit 1406
To determine whether or not it is cached in another space.
If the determination unit 1406 does not exist in another space, that is, if it is cached only in its own space, the input / output processing is performed on the cache block, the lock is released, and the processing is terminated. If the determination unit 1406 determines that the cache block exists in another space, the cache block existing in the own space is copied to the shared cache area in step 1407. In the copy processing at this time, since both the copy source and the copy destination are accessible, there is no need to activate the privileged task.

If the request does not exist in the own space, the judgment unit 1405 judges whether the request is a write request or a read request. If it is a write request, a copy process is performed in the shared cache area in step 1409. In the copy processing at this time, step 8 in FIG.
Copying is performed using a privileged task in the same manner as in step 03. If the determination unit 1408 determines a read request, step 141
If 0, the copy processing is performed to the dedicated cache area in the own space.
This copy process is also performed by using a privileged task as in step 1407.

FIG. 15 shows a modification of the procedure performed by the cache control unit shown in FIG. 9 in the method using the privileged task described in the first embodiment in order to realize the copying means in this embodiment. . In step 1501, a change is made so as to specify the copy destination area when the privileged task is activated, and the determination unit 1505 determines whether the copy has been correctly performed on the specified copy destination. In this case, there is a possibility that the privileged task has terminated abnormally.
In step 1506, the disk block is read into the shared cache area by itself using the replacement means, cached, and input / output processing is performed in step 1506.

FIG. 16 shows a modification of the processing procedure on the privileged task side in FIG. 10 for this embodiment. The privileged task copies to the copy destination area specified at the time of startup in step 1601, and the determination unit 1602 determines the specified area. If the designated area is a shared area, a shared cache block address is set in the entry of the cache block management table in step 1603. If the designated area is a dedicated area, step 1
At 604, an entry in the cache block management table is added to the dedicated cache block address list.
After the above processing is completed, in step 1605, the requesting task is notified that the processing has been completed. By making the above modifications, the copy means of this embodiment is realized.

FIG. 17 shows a flow of processing performed by the replacing means in this embodiment. First, at step 1701, LRU
The last entry in the queue is selected, and in the subsequent processing, the entry in the cache block management table corresponding to the LRU entry is deleted. The determination unit 1702 determines whether the replaced area is a shared area or a dedicated area. If it is a shared cache area, in step 1703, the LRU entries of all the dedicated cache areas included in the entries of the cache block management table are invalidated. Here, an invalid entry is an LRU queue entry for which there is no corresponding cache block management table entry, and indicates that the entry is invalid by substituting 0 into a pointer to the cache block management table entry. When the LRU entry is invalidated for all the dedicated cache areas holding the cache block, the entry in the cache block management table is deleted in step 1704. Judgment unit 17
If the determination at 02 is a dedicated cache area, the determination unit 1705 determines whether the LRU entry is invalid. If not, the address of the dedicated cache area is determined from the entry of the corresponding cache block management table in step 1706 if not invalid. Delete information.

When the release of the cache block to be replaced is completed in the above processing, a data block is read into the cache block in step 1707, an entry in the cache block management table is added in step 1708, and the information of the LRU entry is added in step 1709. Update and move to head of queue.

In this embodiment, replacement of data to be shared is less likely to occur than in the first embodiment in order to further restrict the use of a shared cache area for caching data blocks. As described above, by implementing the cache control unit of the second embodiment, it is possible to further reduce the influence on the cache in the shared area. This embodiment is particularly effective in a system in which a write request hardly occurs as compared with a read request.

[0038]

As described above, according to the present invention, in a system having a cache for improving the access performance to the auxiliary storage device, when sharing does not occur, the sharing used by another application program is not required. Eviction of the area cache can be prevented. Also,
The same data block can be shared by a plurality of processing units.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating a configuration related to cache control of an embodiment in a computer system.

FIG. 2 is a diagram illustrating a process of generating a user space.

FIG. 3 is a diagram illustrating an example of an entry of a cache area management table.

FIG. 4 is a diagram for explaining functions related to input / output processing;

FIG. 5 is a diagram illustrating an example of an entry of a cache block management table.

FIG. 6 is a diagram illustrating an example of an entry of an LRU queue held in each cache area.

FIG. 7 is a diagram illustrating a flowchart of processing performed by a cache control unit in input / output processing.

8 is a diagram showing a flowchart of a partial process when the data block requested in FIG. 7 already exists in the main storage.

FIG. 9 is a diagram showing a flowchart of a partial process performed by a requesting task in order to realize a copy unit.

FIG. 10 is a diagram showing a flowchart of a partial process performed by a privileged task to realize a copy unit.

FIG. 11 is a diagram illustrating a flowchart of a process performed by a replacement unit provided in the cache control unit.

FIG. 12 is a diagram illustrating an example of an entry of a cache block management table installed in a cache control unit according to the second embodiment.

13 is a diagram illustrating an example of a cache block address list included in an entry of the cache block management table illustrated in FIG. 12;

FIG. 14 illustrates processing performed by a cache control unit according to the second embodiment;
FIG. 9 is a diagram showing a flowchart of a partial process when a requested data block already exists in main storage.

FIG. 15 illustrates processing performed by the cache control unit according to the second embodiment;
FIG. 9 is a diagram showing a flowchart of a partial process performed by a requesting task to realize a copy unit.

FIG. 16 is a diagram illustrating a flowchart of a partial process performed by a privileged task to realize a copy unit in the cache control unit according to the second embodiment.

FIG. 17 is a diagram illustrating a flowchart of a process performed by a replacement unit provided in the cache control unit according to the second embodiment.

FIG. 18 is a diagram illustrating a configuration of a computer system including a conventional cache control unit.

[Explanation of symbols]

 101 Main storage device 102, 215, 403 Auxiliary storage device 103 System space 104, 204, 402 Application program 105, 201, 203, 401 User space 106 System call 107, 206, 405 Cache control unit 108, 412 Shared cache area 109, 205 Space control unit 110 Task control unit 111, 211 Cache area management table 112, 407 Cache block management table 113, 406 Search unit 114, 410 Copy unit 115, 207, 411, 415 Dedicated cache area 202 Space start / end system call 208 Cache block area 209, 413 LRU queue 210 Area registration means 212 Sweeping means 214 Registration deletion means 404 Input / output system call 4 8 area selection means 409 replaces unit 414 output unit 416 privileged task control unit

 ──────────────────────────────────────────────────続 き Continuing from the front page (72) Inventor Tsuneo Iida 5030 Totsukacho, Totsuka-ku, Yokohama-shi, Kanagawa Prefecture Inside the Software Division, Hitachi, Ltd. (72) Inventor Hiroshi Sakagami 5030 Totsukacho, Totsuka-ku, Yokohama-shi, Kanagawa Prefecture Hitachi, Ltd. Software Division (72) Inventor Hiroshi Fukuoka 5030 Totsuka-cho, Totsuka-ku, Yokohama-shi, Kanagawa Prefecture F-term in Hitachi Software Division Software Division (Reference) 5B005 JJ11 JJ13 JJ23 KK22 LL03 LL04 MM03 MM05 UU32 5B082 FA12

Claims (4)

[Claims] 1. A cache control method in a computer system comprising an auxiliary storage device, wherein a main storage device is used separately in a system space in which an operating system operates and a user space in which an application program operates, wherein an application is stored in the user space. A dedicated cache area for executing a program is secured, a shared cache area is secured in the system space, and when an application program accesses the auxiliary storage device, data in the auxiliary storage device is temporarily stored in the dedicated cache area. When an access request is issued from another application program to the same data as the temporarily stored dedicated cache area data, the temporarily stored dedicated cache area data is stored in the shared cache area. Ko A cache control method characterized by: 2. The cache control method according to claim 1, wherein an access request is issued from a plurality of application programs to the same data as the temporarily stored data in the dedicated cache area, and all of the access requests are generated. Is a read request, the data of the temporarily stored dedicated cache area is copied to each of the dedicated cache areas for executing the other application programs. 3. A computer system comprising an auxiliary storage device, wherein a main storage device is used separately from a system space in which an operating system operates and a user space in which an application program operates. Means for securing a dedicated cache area, means for securing a shared cache area in the system space, and when an application program accesses the auxiliary storage device, data read from the auxiliary storage device is temporarily stored in the dedicated cache area. Means for storing, when an access request is issued from another application program to the same data as the temporarily stored data of the dedicated cache area, the temporarily stored data of the dedicated cache area is shared. Cap A computer system having means for copying to a storage area. 4. The computer system according to claim 3, wherein an access request is issued from another plurality of application programs to the same data as the temporarily stored data in the dedicated cache area, and all of the access requests are generated. In the case of a read request, a computer system having means for copying the temporarily stored data in the dedicated cache area to each of the dedicated cache areas for executing the other application programs.